Date post: | 14-Jun-2015 |
Category: |
Internet |
Upload: | antonio-alberti |
View: | 389 times |
Download: | 1 times |
© Antônio M. Alberti 2014
INTERNET: CURRENT STATE AND FUTURE ARCHITECTURES - INTRODUCING
NOVAGENESIS PROJECT
Antonio Marcos Alberti !28th August 2014
© Antônio M. Alberti 2014
Outline1. Panorama on the Current Internet State 2. The Race for Future Internet Architectures 3. Examples of Clean Slate Architectures 4. NovaGenesis: Convergent Information Architecture
© Antônio M. Alberti 2014
1. Panorama on the Current Internet Statusü The Internet has invaded most aspects of life and society,
changing our lifestyle, work, communication, and social interaction. !
ü Nobody doubts about the fundamental role of the Internet on our information society. !!!!!
ü The Internet has strong impact on economy, specially on the services sector (70% of GDP on modern economies).
© Antônio M. Alberti 2014
1. Panorama on the Current Internet Statusü However, it was designed in an era where technological
development was completely different from today. !
ü There was not enough capacity to support sophisticated networking services - the solution was to design a simple, but robust network. !
ü The terminals were fixed, inside secure university/government environment - there was not attackers!! !
ü During decades, it was incrementally developed and deployed, achieving impressive scales!!!
© Antônio M. Alberti 2014
1. Panorama on the Current Internet Statusü The result is a complex agglomerate of incremental protocols
that inherits the grown legacies of decades of patchwork solutions. !
ü New protocols must live with the limitations of the preceding. !
ü Intermediary layers have been added to overcome unplanned situations, reducing network efficiency. !
ü While at the physical layer we are approaching to the theoretic Shannon’s limit of channels, at network level, lots of bytes are lost on inefficient stacking. !
ü Let’s consider an example:
Simple
(c) Marco Casaroli 2014
Virtual LAN (VLAN)
(c) Marco Casaroli 2014
Provider Backbone Bridge (PBB)
(c) Marco Casaroli 2014
Multiprotocol Label Switching (MPLS)
(c) Marco Casaroli 2014
Tunnel IP over IP
(c) Marco Casaroli 2014
Virtual Extensible LAN (VXLAN)
(c) Marco Casaroli 2014
V$#^!X!LANWaste of the precious bit rate achieved by physical layer!
(c) Marco Casaroli 2014
Too much control, too few useful information!
© Antônio M. Alberti 2014
1. Panorama on the Current Internet Statusü The Internet looks like a TetrisTM game in which the player is almost "Dead" !!!
IP
TCPUDP
SIP
Overlays
Internet TCP/IP: GAME OVER!?
ATMSDH
© Antônio M. Alberti 2014
1. Panorama on the Current Internet Statusü Other limitations are:
ú The dual semantics of IP addresses (identifier and locator). ú The lack of unique addresses and transparency. ú The inefficient support for host mobility. ú The inexistent support for service mobility. ú The incomplete naming approach, which do not name applications
and other important entities in architecture. ú The weak support for security, privacy, and trust. ú The weak support for multicast. ú The weak support for quality of service. ú The problematic support for inter autonomous system routing. ú The inexistent support for multihoming. ú The scalability issues of client/server model.
© Antônio M. Alberti 2014
1. Panorama on the Current Internet Statusü These limitations motivated many people to question the
adequacy of the current Internet architecture to meet the exponential growths expected to the Internet in the next decades! !
ü Since this question was first made, a lot of initiatives to reshape the Internet appeared around the world - the so called Future Internet design.
© Antônio M. Alberti 2014
1. Panorama on the Current Internet Statusü The main motto question is:
ú Considering the current state-of-art on computing and communications, is it possible to design right now a new Internet that best meets our information society needs?
2013
After redesign!!!
1970
The Beetle
The Internet
No deep redesign until now!!!
Huge scale!
2. The Race for Future Internet Architectures
© Antônio M. Alberti 2014
‣ Map of world wide initiatives
Japão: NICT, Univ. Tokio. -Akari, Net. Virtual. Lab. -JGN, JGN2, JGN2+ -Spark
Map: Wikipédia.
© Antônio M. Alberti
Europa: FP6/FP7/Horizon2020, PPP, KIC, Eureka, EuroNF - EIFFEL, GÉANT2, OneLab & PanLab, NESSI, NEM, RNRT
- 4WARD, SENSEI, S-CUBE, 2020 3D, eMobility, G-Lab, AutoI - CASCADAS, SAIL, IRATI, ABNO, NFV - FIA: FCN, FIRE, SOFI, RWI; FI-PPP: FI-WARE, FINEST - ETPs: NEM, NESSI, Net!Works, EPoS, ENIAC.EUA: NeTS e Outros
- NewArch, 100x100 - FIND, NOSS, ProWin - NBD, GENI, PlanetLab - Internet2 - RINA, XIA, MobilityFirst - CCNx, NDN
Canadá: - Canarie
Ásia: - AsiaFI - CJK
Korea: - FIF, MOFI, - IDCOM, DIANA
Brasil: - RNP, FIBRE - Horizon - RouteFlow - NovaGenesis
China: - CNGI
© Antônio M. Alberti 2014
2. The Race for Future Internet Architecturesü U.S.A.
ü First initiatives: ü RINA (Late 1990-), NewArch (2000-2003): FARA & NIRA,
100x100 Clean Slate Project (2003-2006), Ethane (2006-?), SDN & OpenFlow (2008-), 4D, I3, NSF FIND (2006-), GENI (2005-), POMI, NetFPGA. !
ü More recent initiatives: ü NSF NeTS-FIA: XIA, MobilityFirst, NDN, Nebula, ChoiceNet.
© Antônio M. Alberti 2014
2. The Race for Future Internet Architecturesü Europe
ü First initiatives: ü FP6-IST:
ü Internet of the Future: Ambient Networks (2004-2007), DAIDALOS II (2003-2008) !
ü Future and Emerging Technologies: ANA (2006-2009), HAGGLE (2006-2010), BIONETS (2006-2010), CASCADAS (2005-2007).
!ü More recent initiatives:
ü FP7: ü Hundreds of initiatives: 4WARD, SAIL, TRILOGY, AutoI,
PSIRP, SENSEI, AWISSINET, FIRE, FIBRE, THINK TRUST, E3, SOA4ALL, S-CUBE, NEXOFRA, ASPIRE.
© Antônio M. Alberti 2014
2. The Race for Future Internet Architecturesü Asia
ü Initiatives: ü Japan: Akari, NWGN Forum, JGN2 ü Korea: MOFI, DIANA, IDCOM, FI Forum, KOREN ü China: TUNOS, FINE, SAVA
© Antônio M. Alberti 2014
2. The Race for Future Internet Architecturesü South America
ü Initiatives: ü Brazil: Horizon, NovaGenesis, RouteFlow, FIBRE
© Antônio M. Alberti 2014
2. The Race for Future Internet Architecturesü Stanford University Project Objective:
ü “To reinvent the internet so Future Internet can: ü overcome fundamental architectural limitations of today's
Internet ü incorporate new technologies ü enable new class of applications and services ü continue to be a platform for innovations and thus be an engine
for economic growth and prosperity for the society”. !
ü The 100x100 Clean Slate Project Motto Question: ü “Given the benefit of hindsight and our current understanding of
network requirements and technologies, if we were not bound by existing design decisions and would be able to design the network from first principles (a clean slate design), how should we do it?”
© Antônio M. Alberti 2014
2. The Race for Future Internet Architecturesü The BLED declaration proposed by European Commission:
The BLED Declaration:Towards a European approach to
the Future InternetCurrent Internet: Success & Challenges With over a billion users world-wide, the current Internet is a great success – a global integrated communications infrastructure and service platform underpinning the fabric of the European economyand European society in general. However, today's Internet was designed in the 1970s for purposes thatbear little resemblance to current and foreseen usage scenarios. Mismatches between original design goalsand current utilisation are now beginning to hamper the Internet’s potential. A large number of challengesin the realms of technology, business, society and governance have to be overcome if the future development ofthe Internet is to sustain the networked society of tomorrow.
Future Internet: Vital to continued economic Growth in EuropeIn the future, even more users, objects, services and critical information infrastructures will be networkedthrough the Future Internet which will underpin an ever larger share of our modern and global econo-mies.It is therefore time to strengthen and focus European activities on the Future Internet to maintainEurope’s competitiveness in the global marketplace.
Future Internet: Addressing the Challenges through EU Collaboration & CooperationEU member states have already committed, through the renewed Lisbon Agenda and the i2010 initiative,€9.1 billion of funding, as part of a public-private partnership, for ICT research over the duration of FP7.However, we must ensure that, within this, continuous and long term support is given to the design of theFuture Internet as a key element of the future networked society. It is of strategic importance for Europeto fully engage in the conception, development and innovation of a Future Internet ensuring the longterm growth of the ICT sector, full support to an ICT based economy, and the elimination of the digitaldivide for all citizens.The research projects assembled here in Bled represent the first phase of this public-private partnership,a joint investment of over €400 million, that recognises the challenges above and emphasises a concer-ted and comprehensive process of redesign, based upon novel network, service, trust, security and contenttechnologies together with strong initiatives towards new innovations in societal, governance and servicedomains, in order to ensure that the Future Internet fulfils its potential.More specifically, building upon the obligations of our individual project contracts and the goals of theStrategic Agendas of the European Technology Platforms, we confirm our ambitions include:Fostering Favourable Conditions through Coordinated Action•• Coordinate our efforts to foster cross-disciplinary innovation and creativity.•• Work together through a European Future Internet Assembly of research projects strengthening
cross-discipline activity and optimising the impact of our actions.•• Cultivate and foster the skills and knowledge required to develop the Future Internet.•• Create the conditions for the deployment of services and service oriented systems.•• Communicate through open standards for Future Internet technologies and architectures.•• Open the European Future Internet Assembly to new projects and actors over time to widen the
coordination and consistency of the action.Jointly Designing, Developing and Experimenting•• Services and networking architecture for the Future Internet.•• Location independent, interoperable, coherent, consistent, scalable, pervasive, reliable, secure and
efficient access to a coordinated set of services.•• Tools supporting collaborative business models and social network applications.•• Technologies ensuring the robustness and security of the networks, managing identities, protecting
privacy and creating trust in the on-line world.•• Approaches and tools to leverage the full potential of the Internet of Things.•• Capabilities for supporting the creation, sharing, locating and delivery of new-media content.Increasing Awareness at Policy Level•• Raise awareness of the economic, policy and regulatory issues as identified by the newly proposed
European Future Internet Assembly, the UN Internet Governance Forum, the OECD and the European regulatory frameworks.
•• Contribute to the definition of European positions within global forums and arenas.
Call for European action towards the Future InternetTo help us meet these major challenges, we call on the:•• European Member States to strengthen and coordinate their national R&D efforts and initiatives
toward the Future Internet.•• European Commission to stress the vision and amplify the related R&D in order to drive Europe ahead
of tomorrow’s Internet transformations in the way we work, live, and interact.•• European Member States and the European Commission to support the creation and activities of the
European Future Internet Assembly proposed in this declaration.
This declaration is endorsed by the following European Technology Platforms andEuropean Research Projects*: eMobility, NEM, NESSI, ISI and EPOSS2020 3D Media 4NEM 4WARD ADAMANTIUM AGAVE ASPIRE AUTOI AVANTSSAR AWISSENET CASAGRAS CHIANTI
CHORUS COIN CONTENTCuteLoop DICONETE3 eCRYPT II EFIPSANS EIFFEL eMOBILITY EURO-NF
FAST FORWARDINTERSECTION IRMOS iSURF m CIUDAD MASTER MobileWeb2.0 MOBITHIN MOMENT NAPA-WINE
N-CRAVENESSI 2010 OPEN P2P NEXT PanLab / PII PERSISTPetaMedia PICOS PRIMELIFEPRISM RESERVOIR
SAPIR S-CUBE SEA SENDORASENSEI SERVFACE Service WEB 3.0SHAPE sISI SMOOTH-IT SOA4ALL
SOCRATES SWIFT TA2TAS3 TECOM THINK-TRUST VICTORY WOMBAT
A significant change is required and the European Internet scientific and economicactors, researchers, industrialists, SMEs, users, service and content providers, now assert the urgent necessity to redesign the Internet, taking a broad multidisciplinaryapproach, to meet Europe’s societal and commercial ambitions.
*Accession to this declaration is open to existing and future EU Projects that wish to actively contribute
153-08 The Future Internet Manifesto:Mise en page 1 17/03/08 10:34 Page 1
EU member states have already committed, through the renewed Lisbon Agenda and the i2010 initiative, €9.1 billion of funding, as part of a public-private partnership, for ICT research over the duration of FP7.
3. Examples of Clean Slate Architectures
© Antônio M. Alberti 2014
‣ RINA: Recursive Internet Architecture
‣ XIA: eXpressiveness Internet architecture
‣ NovaGenesis
© Antônio M. Alberti 2014
3. Examples of Clean Slate Architecturesü RINA:
ü Designed considering that networking is IPC (Interprocess Communication), and only IPC. !
ü There are four RINA implementations: ü Alba ü TRIA network systems LLC ü Boston university ü Investigating RINA as an Alternative to TCP/IP (IRATI) funded
by FP7 in Europe.
© Antônio M. Alberti 2014
3. Examples of Clean Slate Architecturesü RINA:
ü The IRATI`s key RINA abstractions are: ü Distributed IPC facility (DIF) ü Distributed Application Facility (DAF) ü Resource Information Base (RIB) ü Inter-DIF Directory (IDD) ü Error and Flow Control Protocol (EFCP) ü Common Distributed Application Protocol (CDAP) ü Common Application Connection Establishment Phase
(CACEP).
Layers
Network
Transport
Data Link
Physical
Applications
Network
Transport
Data Link
Physical
Applications
Network
DL DL
PHY PHY
TCP, UDP, …
IP
Web, email, ftp, …
RINA
Network
Transport
Data Link
Physical
Applications
Network
Transport
Data Link
Physical
Applications
Network
DL DL
PHY PHY
TCP, UDP, …
IP
Web, email, ftp, …
DIF – “Networking is IPC”
Network
Transport
Data Link
Physical
Applications
Network
Transport
Data Link
Physical
Applications
Network
DL DL
PHY PHY
TCP, UDP, …
IP
Web, email, ftp, …
IPC
IPC IPC
ü All DIFs have the same interface and components, regardless of their architectural level. DIFs are populated by IPC processes, which have a common internal structure.
An IPC Process
IPC API
Data Transfer Data Transfer Control Layer Management
SDU Delimiting
Data Transfer
Relaying and Multiplexing
SDU Protection
Transmission Control
Retransmission Control
Flow Control
RIB Daemon
RIB CDAP Parser/Generator
CACEP !!Enrollment
Flow Allocation
Resource Allocation
Forwarding Table Generator
Authentication
State VectorState VectorState Vector
Data Transfer Data Transfer
Transmission Control
Transmission Control
Retransmission Control
Retransmission Control
Flow ControlFlow Control
Increasing timescale (functions performed less often) and complexity
© Antônio M. Alberti 2014
3. Examples of Clean Slate Architecturesü XIA:
ü A key abstraction is the principal, which can be defined as some type of entity that deserves attention, e.g. hosts, services, content.
!ü The key XIA concepts are:
ü Unique identification of every principal ü Expressing intent ü Flexible addressing using unique IDs ü Iterative refinement ü Intrinsic security !
ü Self-certifying identifiers (XIDs) are calculated by hashing the public cryptographic keys of domains, hosts, and services or the entire binary pattern of contents.
© Antônio M. Alberti 2014
3. Examples of Clean Slate Architecturesü XIA:
ü To express intent means to give clues to the network - XIDs - that point clearly the desired communication targets (e.g. a content, a service, a host, or a domain). !
ü Provides a flexible addressing approach where alternative forms of accessing some communicating target (called fallbacks) are provided. !
ü XIA routers can use more than one destination address to forward packets to a desired target. !
ü Therefore, if some ID is unknown at some router, alternative addressing can help to found the desired communicating target (a Content ID for example).
© Antônio M. Alberti 2014
3. Examples of Clean Slate Architecturesü XIA:
ü Proposes: ü A principal-independent expressive Internet protocol (XIP) that
covers flexible addressing, packet format, routing based on XIDs, and in-network caching of data chunks. !
ü A socket API - similar to the UDP socket. !
ü Three transport protocols: • X-datagram protocol (XDP), X-stream protocol (XSP), X-
chunk protocol (XChunkP).
4. NovaGenesis: Convergent Information Architecture
© Antônio M. Alberti 2014
‣ 2008: Started Future Internet activities @ Inatel.
‣ 2009: Studying.
‣ 2010: First NovaGenesis design principles and choices.
‣ 2011: Finished draft architecture.
‣ 2012: Started proof-of-concept implementation. Refinement of choices.
‣ 2013: Test scenario over Mobile Future Internet (MOFI) from ETRI, CNU, and KNU - South Korea.
Very Brief History
© Antônio M. Alberti 2014
‣ What is it?
‣ A Clean Slate Convergent Information Architecture
‣ A Framework Easily Integrated with the Current Internet
‣ An Alternative to the Current Internet Architecture
NovaGenesis
subscribing of services require authentication, secrecy, integrity, and authorization. Trust networks can be created among services in a service cluster. The same challenge reaches RWI. NEDs require innovative security mechanisms, including trust net-works. Since localization, context, and objective information are sensible, generalized mobility, semantic, and context also requires SPT support.
Real-time support is necessary for multimedia applications; to enable timely feed-back control on autonomic loops, to enable virtual networks to quickly react to changes on substrate resources, and to provide freshness of information on RWI. In addition, virtual entities, e.g., virtual base stations, can also require real-time support to promptly react to environmental changes. Reconfigurable hardware is important for virtualization and experiment-driven research. Nonetheless, virtualization is required to provide evolvability and sustainability of reconfigurable hardware.
Service-based applications accommodate compose-ability, diversity, extendability, and openness of applications, but requires autonomicity, security, privacy, context,
Fig. 1. Full mesh of synergies and dependencies among future Internet ingredients.© Antônio M. Alberti 2014
The NovaGenesis Team
(c) A
nton
io A
lber
ti 20
13, I
nate
l.
NovaGenesis in the Media
(c) A
nton
io A
lber
ti 20
13, I
nate
l.
© Antônio M. Alberti 2014
Papers
(c) A
nton
io A
lber
ti 20
13, I
nate
l.
© Antônio M. Alberti 2014
‣ Adopted decision choices:
- Entities and content naming using natural language and self-certifiable names (hash codes).
- As functionalities are seen as services, including network protocols implementations.
- Complex protocols like TCP are fragmented on a population of cooperating services - combined at runtime.
- Name bindings are stored on distributed hash tables, representing all kind of relationships among named-things.
- Name bindings are published and subscribed, enabling distributed search, discovery, negotiation, and contracting of services and content.
Overview
‣ Adopted decision choices (cont.):
- Substrate resources are exposed to software by proxies, which represent them regarding resource life-cycling and orchestration.
- All the communication is done by name-based message scheduling and exchanging, with dynamic headers.
- All the contracts among services can capture intrinsically the required quality, security, privacy, reputation, etc.
- The services will employ a decision cycle to meet objectives traced by human and machine operators.
- They will compete each other to better satisfy contracts (evolutionary pressures) and optimize the usage of substrate resources (evolution environment).
Overview
NovaGenesis Overview
© Antônio M. Alberti 2014
Testing Scenario: Joint Services and Content Orchestration
‣ Clients e servers discover each other, publish SLAs and exchange named-content based on user policies.
‣ Provides check for content provenance, integrity, and coherence of content.
‣ MOFI provided transparent connectivity change.
© Antônio M. Alberti 2014
‣ Content publishing round trip time
!!!!!‣ Content subscription round trip time
Testing Scenario: Joint Services and Content Orchestration
© Antônio M. Alberti 2014
‣ NovaGenesis as an architecture to Adaptive and Cognitive Radio over Fiber (ACRoF) and Internet of Things (IoT)
From “H”RoF RoF
Splitter
Access Point
Spectrum Analyzer
Antenna Control Link
From “H”
To “E” To “E”
Optical Switch
Throughtput (Mbps)CINR (dB)
RF
Sa Freque © Antônio M. Alberti 2014
NovaGenesis Services for Proxy/Gateway/Control of: !-Spectrum Sensing -Optically
Controlled Antenna -Access Point -Wi-Fi VLAN !!!
‣ NG protocols are implemented as services, thus they have the same structure and life-cycle than any other service.
‣ NG creates a dynamic stack, where protocol implementations can be replaced any time and used on more than one layer, simultaneously.
‣ It brakes RINA IPC process structure or XIA protocol implementations in a set of dynamically composed services.
‣ NG enables any service to be particularized to any kind of entity.
‣ There is no limit on the number of principals or on what networking services can be customized to a principal type.
‣ NG also employs self-certifying names, but it enables any service to use natural language names, as well.
NG advantages when compared to RINA or XIA
‣ NG core services are available to applications via dynamic contracting, releasing them of duplicating the same services at upper layers.
‣ NG considers joint content and service life-cycling as a fundamental requirement.
‣ NG includes search, discovery, and reputation of services in the core.
‣ NG supports any naming structure, in any language.
‣ NG provides a distributed, scalable, generic name resolution service.
‣ NG avoids the RINA inter-DIF complications, and generalizes and expands XIA expressiveness.
NG advantages when compared to RINA or XIA
© Antônio M. Alberti 2014
Thank You!
Antônio Marcos Alberti !
antonioalberti.blogspot.com facebook.com/antoniomarcos.alberti
researchgate.net/profile/Antonio_Alberti linkedin.com/profile/view?id=69752898
http://inatel.academia.edu/AntonioMarcosAlberti mendeley.com/profiles/antonio-marcos-alberti/
twitter.com/antoniomalberti